Hermes Agent Bedrock support lets you run Hermes on models hosted in your own AWS account — Claude, Nova, DeepSeek, Llama and more — through Amazon Bedrock. Hermes uses Bedrock’s Converse API, so one setup works across model families, and it uses your normal AWS credentials. In the video below I cover a new model landing on Bedrock. Here’s how to connect Hermes to Bedrock step by step.
Key takeaways
- Hermes has a built-in bedrock provider.
- It uses the Converse API — one config for Claude, Nova, DeepSeek and Llama.
- Auth uses the standard AWS credential chain — no extra Hermes key.
- Add Bedrock Guardrails in config if your company needs them.
Why use Hermes Agent with Bedrock
- Your AWS account — billing, limits and data handling stay inside AWS.
- Company compliance — many businesses already approve Bedrock.
- One provider, many models — Claude, Nova, DeepSeek, Llama and others.
- Guardrails — apply Bedrock Guardrails to agent output.
- Cross-region inference for availability.
Hermes Agent Bedrock setup
The easy way: run hermes model and choose “AWS Bedrock”.
From the command line:
hermes chat --provider bedrock --model us.anthropic.claude-sonnet-4-6With a specific AWS profile and region:
AWS_PROFILE=myprofile AWS_REGION=us-east-1 hermes chat --provider bedrock --model us.anthropic.claude-sonnet-4-6Swap in any Bedrock model ID you have access to.
Want this working in your business, not just bookmarked? Running AI agents on your company’s cloud is exactly what we build together inside the AI Profit Boardroom — 3,400+ members, four live calls a week, and a full one-hour DeepSeek Harness course in the classroom.
Prefer it mapped 1-on-1 first? Book a free strategy session and we’ll plan it for your exact situation.
Join the AI Profit Boardroom →Book a Free Strategy Session →
Hermes Agent Bedrock config.yaml
To make Bedrock your default, add this to your Hermes config:
model:
provider: "bedrock"
default: "us.anthropic.claude-sonnet-4-6"
bedrock:
region: "us-east-1"
# profile: "myprofile"
# guardrail:
# guardrail_identifier: "your-guardrail-id"
# guardrail_version: "DRAFT"Set BEDROCK_BASE_URL only if you need a non-default regional endpoint.
AWS credentials for Hermes
Hermes uses the standard AWS credential chain, so if the AWS CLI already works on your machine, Hermes will too. That covers access keys, a named profile in ~/.aws/credentials, IAM roles, instance metadata on EC2, and SSO. Your IAM user or role needs permission to invoke the Bedrock models you want, and model access must be enabled in the Bedrock console for your region.
Choosing a Bedrock model for Hermes
| Need | Model family to try |
|---|---|
| Hard reasoning and coding | Anthropic Claude |
| Low cost at scale | Amazon Nova |
| Open-weight reasoning | DeepSeek |
| Open models | Meta Llama |
Agents need plenty of context — Hermes needs at least 64K — so pick models with large context windows.
Hermes Agent Bedrock troubleshooting
- Access denied — enable model access in the Bedrock console and check IAM permissions.
- Model not found — check the model ID and that it’s available in your region.
- Wrong account — set AWS_PROFILE explicitly.
- Slow or throttled — try cross-region inference or request a quota increase.
When Bedrock is the right choice
Bedrock is ideal if your business already runs on AWS. Billing goes through your existing account, IAM controls who can use which models, and many companies have already approved Bedrock for sensitive work. It also gives you several model families through one provider, so you can switch from Claude to Nova or DeepSeek without changing your setup. For a solo user without an AWS account, though, it’s usually simpler to start with Nous Portal or OpenRouter.
You can also mix providers: Bedrock for client work that must stay in AWS, and a free Nous Portal model for internal research.
Bedrock IAM checklist
Before connecting Hermes, make sure the IAM user or role it uses can invoke the Bedrock models you want, and that model access is enabled for those models in the Bedrock console in your region. Use a dedicated role or profile for Hermes rather than your personal admin credentials, give it only the Bedrock permissions it needs, and add Bedrock Guardrails if your company requires content filtering. If you use cross-region inference, check the model IDs include the right region prefix.
Using Bedrock with Hermes skills and cron
Once Bedrock is connected, everything else in Hermes works as normal: your skills, memory, tools and scheduled jobs simply use the Bedrock model as their brain. That means you can run a daily research job or a weekly report entirely inside your AWS account, which is often exactly what a compliance team wants to hear.
Frequently asked: Hermes on Bedrock
Is Bedrock more expensive than other providers? It depends on the model; you pay AWS’s Bedrock rates for each model, billed to your AWS account. Can I use Bedrock and Nous Portal together? Yes — set up separate Hermes profiles and switch between them. Does Hermes send my data anywhere else? Model calls go to Bedrock in the region you choose; your memory and files stay wherever Hermes runs. Can I restrict which models the team uses? Yes, through IAM policies and Bedrock model access settings.
The bottom line on Hermes Agent Bedrock
Hermes Agent Bedrock setup takes a few minutes: choose AWS Bedrock in hermes model, use your normal AWS credentials, and pick a model. It’s the best route if your business already runs on AWS.
If you’re the person setting this up for a team, document the model IDs, region and IAM role you chose, and share a short guide so everyone connects the same way. Hermes profiles make this easy: create a Bedrock profile once, export it with hermes profile export, and share it without credentials so each person adds their own AWS access. That keeps setups consistent and audit-friendly while letting each user work in their own account or role. Finally, monitor usage in the AWS console for the first few weeks. Agents can make many model calls, especially with sub-agents and scheduled jobs, so set a budget alert in AWS Billing and choose cheaper Bedrock models for frequent background tasks. That keeps costs predictable while you find the right balance of quality and price.
FAQ: hermes agent bedrock
Does Hermes Agent support AWS Bedrock?
Yes — with the built-in bedrock provider, using Bedrock’s Converse API.
How do I connect Hermes to Bedrock?
Run hermes model and choose AWS Bedrock, or use hermes chat –provider bedrock –model followed by a model ID.
Do I need a special Hermes key for Bedrock?
No — Hermes uses your standard AWS credentials.
Which models can Hermes use on Bedrock?
Claude, Nova, DeepSeek, Llama and other models you have access to in your region.
Can I use Bedrock Guardrails with Hermes?
Yes — add a guardrail identifier and version under bedrock in config.yaml.
Two ways I can help from here. Join the AI Profit Boardroom — it’s where your Hermes model setup gets built with 3,400+ members doing the same.
Or grab a free strategy session and bring your questions.
Join the AI Profit Boardroom →Book a Free Strategy Session →
About Julian Goldie
I’m Julian Goldie — founder of Goldie Agency, best-selling author, and an AI educator with 428K+ YouTube subscribers. I’ve spent 10+ years in SEO and online business and I test every AI agent on my own work before I recommend it. Join the AI Profit Boardroom for the daily builds, or book a free strategy session to talk through yours.
Related reading
Last updated October 2026. This page is a living guide to hermes agent bedrock — the facts here move fast and it is updated as they do.